Machine Learning for Soil Fertility and Plant Nutrient Management using Back Propagation Neural Networks

The objective of this paper is to analysis of main soil properties such as organic matter, essential plant nutrients, micronutrient that affects the growth of crops and find out the suitable relationship percentage among those properties using Supervised Learning, Back Propagation Neural Network. Although these parameters can be measured directly, their measurement is difficult and expensive. Back Propagation Networks(BPN) are trained with reference crops’ growth properties available nutrient status and its ability to provide nutrients out of its own reserves and through external applications for crop production in both cases, BPN will find and suggest the correct correlation percentage among those properties. This machine learning system is divided into three steps, first sampling (Different soil with same number of properties with different parameters) second Back Propagation Algorithm and third Weight updating. The performance of the Back Propagation Neural network model will be evaluated using a test data set. Results will show that artificial neural network with certain number of neurons in hidden layer had better performance in predicting soil properties than multivariate regression. In conclusion, the result of this study showed that training is very important in increasing the model accuracy of one region and result in the form of a guide to recognizing soil properties relevant to plant growth and protection.